Updating on News Without Overreacting
New information should move your forecast by an amount proportional to how surprising it is. Most people move too far on noise and not far enough on signal.
A forecast is a live estimate, not a one-time guess. The skill is moving it the right distance when something happens.
The question that governs everything
How much more likely is this news if the thing is true than if it is false?
That ratio is all that matters. If a piece of news is roughly as likely either way, it should not move your forecast at all, regardless of how dramatic it is. If it is five times more likely under one hypothesis, it should move it substantially.
Most news is in the first category and gets treated as though it were in the second.
The rough arithmetic
Work in odds rather than percentages.
Start at 20%, which is odds of 1 to 4. A piece of news three times more likely if true than if false multiplies the odds by three: 3 to 4, which is about 43%.
That is it. Convert to odds, multiply by the likelihood ratio, convert back. Odds formats covers the conversions.
You will not do this precisely in your head and you do not need to. The value is in the habit of asking for the ratio, which automatically deflates news that sounds important and is not.
The two errors
Overreacting to vivid non-evidence. A well-argued opinion piece. A poll within its margin of error. A statement from someone with an obvious incentive. These feel like information and have likelihood ratios close to one.
Underreacting to boring evidence. A procedural deadline passing. A quiet withdrawal. A scheduled step not happening. These are frequently strong evidence and produce no headline.
The asymmetry is systematic: attention-getting news is selected for being interesting, not for being diagnostic.
Watching the market update
One of the genuinely useful things about a live market is that it shows you the aggregate update in real time. News breaks, the price moves, and the size of the move is the market's collective likelihood ratio.
Compare that against your own. If you moved twenty cents and the market moved two, one of you is wrong and it is worth asking which. Top movers is where that shows up on this site.
Be careful about the direction of causation, though: if you look at the price before forming your own update, you have not made an independent judgement. Form yours first. Where crowd wisdom runs out.
When a big move is justified
The event itself. Not news about it — the thing happening or being formally confirmed.
A source with genuine access. Reporting that would be very unlikely if the thing were not happening.
An elimination. In a multi-outcome market, one outcome being ruled out redistributes its probability across the rest, mechanically. Multi-outcome markets.
A deadline passing. Time is evidence. Something that has not happened by the point it usually would have has become less likely, and this update happens continuously whether you make it or not.
The discipline
Write the update down with its reason, and write down what you expected before it happened. Over time you will discover which categories of news you systematically over-weight, which is a specific and fixable defect rather than a vague one. Keeping a forecast journal.
Next
Educational content only. Nothing here is financial advice, a solicitation, or an offer to trade. Prediction markets can move against you and a position can lose its entire value. Where this guide describes a difference between venues it describes a theoretical cross-market edge in observed prices, which is not a profit: fees, spreads, capital lockup, transfer time and differing settlement rules all come out of it first.
Keep reading
Forecasting and Calibration · 3 min readLog Scoring, and Why Confident Errors Cost So MuchThe other proper scoring rule. It is unbounded, it punishes confident mistakes severely, and that severity is the argument for using it.
Forecasting and Calibration · 3 min readThe Habits That Actually Correlate With AccuracyResearch on forecasting tournaments found consistent behavioural differences in the best performers. Most of them are habits rather than talents.
Forecasting and Calibration · 3 min readThe Brier Score, Explained ProperlyThe standard way to score a probabilistic forecast, what a given number actually means, and why it rewards honesty rather than confidence.
Cross-Market Gaps · 3 min readWhy Two Venues Price the Same Event DifferentlyA price gap between two prediction markets is usually not an error. Here are the structural reasons the same event trades at different numbers in different places.